Get our free email newsletter

Basics of EMI Troubleshooting

1403 emi-troubleshooting coverSooner or later, anyone involved with EMI will be involved in troubleshooting an EMI problem, wherever it may surface. Most commonly, the problems will be uncovered during EMI testing, generally very late in the product design cycle, resulting in costly patches and schedule delays. It is best if preliminary EMI testing is done early in the design stage – EMI problems can be uncovered early enough that corrective action can be done in a timely fashion, ideally at the circuit board level. On the back end, EMI problems are often encountered in the field – perhaps because the environment is harsher than that expected by the regulatory agencies or because of an installation problem.

In each of these situations, there are a wide variety of problems that can occur: there may be multiple problems co-existing, and there is usually more than one way to fix the problem. Considering the range of problems, it would seem that EMI troubleshooting is a hit-or-miss situation. Nevertheless, a reasonable methodology can be formulated to minimize the false starts. There will never be a sure fire approach to running down problems, but the process can be minimized.

This article will categorize the basic EMI problems, where they are likely to occur, and the tools available for running the problem to earth.

- Partner Content -

A Dash of Maxwell’s: A Maxwell’s Equations Primer – Part One

Solving Maxwell’s Equations for real-life situations, like predicting the RF emissions from a cell tower, requires more mathematical horsepower than any individual mind can muster. These equations don’t give the scientist or engineer just insight, they are literally the answer to everything RF.


Let’s start by summarizing the problems likely to be encountered, as evidenced by modern EMI test requirements. The US military formulated the basic terminology: emissions and susceptibility, and both radiated and conducted paths: RE, RS, CE, CS. (In more recent times, the term “susceptibility to interference” is commonly replaced by “immunity to interference,”)

So, there are basically two considerations: emissions vs susceptibility and conducted vs radiated. A closer look at these will give clues as to how to proceed.

Problems uncovered during EMI testing are definitive, with specific frequencies and levels being readily available. Problems uncovered in the field are much more elusive, as the cause of the problem may not be obvious. In this case, it will be necessary to identify the cause before effective remedial action can be taken.

Let’s start with a summary of these four basic issues, how they occur, and when they occur.

- From Our Sponsors -

Emissions from electrical and electronic equipment are almost exclusively uncovered during EMI testing, and are perhaps the most common of EMI test failure: the limits are set to prevent interference to sensitive nearby radio receiving equipment. Emissions are generally too low to pose a threat to nearby ordinary electronic equipment. Accordingly, emission problems are rarely encountered in the field.
Since radio receiving equipment operates on continuous waves, the source will also be periodic waves and their harmonic frequencies, most commonly oscillators and switching power devices.

In the field, heavy starting loads and inductive kick from turn-off in your equipment may affect other equipment – these won’t typically be uncovered during EMI testing, another reason for uncertainty in the field.

Immunity (or Susceptibility)
External interference sources assaulting the equipment are varied, as evidenced by the array of susceptibility tests: conducted and radiated RFI, power transients, lightning and ESD, to mention the most common. In addition to radio sources, transients from nearby equipment, notably power loads, become an issue, as does lightning.

This is basically the opposite of emissions (interference getting into the box as opposed to interference getting out of the box) and, as such, the fixes are largely reciprocal.

Conducted or Radiated EMI
Interference may enter or leave the enclosure by conduction via a data or power cable or by radiation, which may be directly through the enclosure or via a data or power cable. The culprit needs to be identified, as it is a necessary path to a solution. To understand why, we first need to understand what facilitates radiation.

Effective radiation requires a suitable antenna to receive or transmit, which requires a metallic element that is a significant fraction of a wavelength. So the first order of business is to establish the wavelength, which is calculated from the offending frequency:

λ = 300/f, where f is frequency in MHz and λ is wavelength in meter.

For continuous waves, frequency is determined by test. For transients, use the bandwidth of the pulse, which is 1/(π*tr). As an example, ESD has a rise time of 1 ns, providing a bandwidth of about 300 MHz.

Having determined the wavelength, look for metallic members greater than about 1/20 wavelength. In actuality, resonances occur at 1/4 or 1/2 wavelength, where radiation is near optimal, so anything approaching 1/4 wavelength becomes significant. This applies to dipole antennas (like cables), slot antennas (openings in metallic enclosures), and loops (internal cables and circuit board traces).

Table 1 gives some representative dimensions as a function of frequency.

Frequency 1/20 1/4 1/2
1 MHz 15 meter 75 meter 150 meter
10 MHz 1.5 meter 7.5 meter 15 meter
30 MHz 50 cm 2.5 meter 5 meter
100 MHz 15 cm 75 cm 1.5 meter
300 MHz 5 cm 25 cm 50 cm
1 GHz 1.5 cm 7.5 cm 15 cm

Table 1: Dimensions for Effective Radiation


The bottom line is, low frequencies don’t radiate effectively, as there are few metallic members large enough to make a good antenna. At 30 MHz, where commercial radiated emission tests start, the only metallic elements long enough to serve as effective antennas are cables. Enclosure dimensions, being much smaller, don’t become a consideration until about 300 MHz.

Alternately, high frequencies don’t conduct well, due to lead inductance in wires and cables, so conducted EMI is largely a low frequency issue. Overall, common radiated RFI frequencies tend to be a few hundreds of MHz, which puts cables as primary suspects. Low frequencies, such as from switched mode power supplies (SMPS) and motor drives, tend to dominate at conducted frequencies, below 30 MHz. Yes, these are rules of thumb – there will certainly be radiated problems below 30 MHz and conducted problems above 30 MHz. Military standards test with a considerable frequency overlap.

Transients vs Continuous Waves
As mentioned above, transients don’t show up in emissions testing. For immunity, both transients and continuous waves may cause problems. Transients tend to create digital errors while continuous waves tend to cause analog input errors.

Further, emissions tend to originate from low impedance circuits such as output drivers and switching power circuits. Immunity tends to attack high impedance circuits, such as op amp inputs and feedback circuits in voltage regulators


All interference problems have three common elements: There is always a source of interference, a receptor of interference and a path linking the source to the receptor. It is usually not possible to eliminate the source or the receptor, so the remaining choice is to attack the path. Table 2 shows some possibilities. Depending on the problem, the source or receptor (or both) may be apparent, but, if not, they will need to be identified.

• Microprocessors
• Video Drivers
• Transmitters
RF Heaters
• Power Disturbances
• Lightning
• Radiated
° EM Fields
° Crosstalk
– Capacitive
– Inductive
• Conducted
° Signal
° Power
° Ground
• Digital
° Microprocessors
° Reset
° Other Logic
• Low Level Analog
• Receivers


Table 2: Interference Involves a Source, a Path, and a Receptor

We have an acronym that we use in identifying the key parameters during the design phase, FAT-ID – frequency – amplitude – time – impedance – dimensions. The acronym is useful for troubleshooting, as well. Once the source and receptor has been identified (perhaps tentatively), the next steps are to identify these parameters:

Frequency – identifying the problem frequencies is the first step in troubleshooting – all remedial steps depend on this information, Lacking this, you are reduced to guessing, and this is not a productive approach. Test results will provide this information but with field problems you may need to hunt, or guess.

Amplitude – what is the amplitude relative to expectations? Is the problem modest, in which case mild fixes may be adequate? Or will major efforts be required?

Time – this can have several aspects. During EMI testing, it may be during a particular operational state of the equipment. If in the field, it may be a particular time of day or season.
Impedance – this will be a factor with I/O and filter design.

Dimensions – depending on the problem frequencies, potential antennas may be found in cable length or enclosure openings.


Problems that surface in the field are almost always harder to run down than test lab problems. In a test lab, the failure problems are specifically identified with calibrated data, and the efficacy of the fixes can be readily evaluated.

In the field, the source of the problem is often unknown, and may well be intermittent: the failure may occur at seemingly random times and there may be no apparent sources. So the problem is to figure out what caused the failure, patch in fixes, then to be confident that the problem has been fixed.

Handling Field Problems – Identifying the Source
The first order of business is to identify the cause of the problem, recognizing that the cause may not be directly found. At the site, look for possible causes. Here are the common possibilities:

Radio sources – Possible sources are nearby radio and TV transmitters, fixed base commercial and emergency transmitters. Internal to the facility, look for RF heaters, arc welders, and use of hand held radios – these are mobile and sporadic. If vehicles may be stopped in a garage, consider on-board radio transmitters.

Power disturbances – look for heavy equipment, large motors, etc., common in industrial facilities and often found in commercial office buildings (elevators and air conditioners, for example). Depending on the geographical area and the season, lightning strikes may be a factor.

Electrostatic DischargeESD can be a problem any time the humidity is low, particularly in the heating season. There may be static generators within the facility, such as conveyer belts and paper or plastic film rolling. In extreme cases, ESD below the human threshold of feeling (about 2 kV) may cause equipment anomalies.

Handling Field Problems – Forcing the Failure
Running down a field problem is nearly impossible without being able to evaluate whether your corrective action is effective. To do this, it is vital to be able to force the failure, which requires some external equipment. Formal test equipment is desirable, but often not available or permissible.

ESD guns are readily available, are portable, and relatively inexpensive. ESD applied directly to the equipment carries some risk of damage, and this can be especially problematic in the field, as the equipment may be in actual operation at the time in question. So if ESD is a possibility, test with extreme caution, starting with indirect discharge if possible, followed by very low level direct contact. Inexpensive ESD sensors are useful in detecting possible ESD sources.

Handheld radios can often be used to identify suspected radio interference. E field can be estimated by: E = 5.5*SQRT(P)/R. For a one watt transmiter, an E field of 10 v/m will be achieved at a distance of about ½ meter. Start by irradiating cables – analog sensor inputs are most vulnerable, followed by power cables. If possible, employ a handheld radio in use at the facility, usually from security or maintenance people. Common radio bands run at about 150 and 450 MHz. Hold the radio parallel to the cables, starting at a distance of maybe two meters, and closing in until a failure is observed. Then proceed to the equipment enclosure itself, repeating the procedure.

Power disturbances are difficult to simulate in the field, due to operating constraints in the field – injection of a transient may adversely affect nearby equipment sharing the same power source. If you have access to a power transient generator, you can proceed much as you would in the test lab. A power quality monitor is useful for identifying transient effects. Connect to the power line and let it run, preferably long enough to observe the failure. Cycling nearby equipment may force a failure, expediting this process.

A chattering relay might be used to inject transients into the power line: the relay coil is connected in series with the normally closed contact – the relay doesn’t know if it should be on or off, so it chatters. It generates copious amounts of noise into the power line, perhaps too much for comfort.

Another possibility is to inject a transient into the power line. Tape an 18 inch length of wire to the power cord, ground one end of the wire and discharge into the other end. This will inject a fast transient into the power cord. Sneak up on the level, much as with the ESD test procedure described above.


We’ll start by assuming you know, or have a reasonable handle on the interference source and recipient. This would definitely be true if you are working with test lab data, but may or may not be true with field problems.

  1. Start with FAT-ID. This will provide initial insight to the nature of the problem. Knowledge of frequency helps you to identify significant metallic members and impedance of critical paths.
  2. Minimize the system. Remove all unneeded cables and power down unneeded equipent. The goal is to start with as few variables as possible. Where cables are necessary, use clamp-on ferrites to minimize cable effects. Establish a baseline failure, compare with that of the unmodified set-up. If there is still a failure, it’s time to work on the system as is – power input is a good place to start.
  3. As improvements occur, remove the ferrites and continue by adding cables. Evaluate and fix as you progress. If your enclosure is non-conductive, you will need to attack at the circuit boards. If accessible, apply fixes at the board boundary, typically with filters.
  4. If you can’t eliminate the problem by working the cables or if the problem persists even with a minimum system, turn to the enclosure. If you have a metallic enclosure, close the seams for effect, using conductive copper tape or wrap in aluminum foil.

Knowing the problem frequencies gives some clues as to where to start. Frequencies below about 30 MHz are usually conducted, often power line related. Above 30 MHz, problems are usually cable related. Above about 300 MHz, enclosures and circuit boards start to become contributors.


First to note, when troubleshooting, there may well be more than one problem, and those may be handled by the same fix, or a combination of fixes. In particular, with a continuous wave frequency range, the same problem may show up at a number of frequencies. Here, it is usually best to start by attacking the lowest frequencies first – often the higher problem frequencies will diminish as well.

Second, the fixes you try during troubleshooting will usually be different than you would have used during the design phase. As a general rule, you design your equipment from the inside out, and you fix it from the outside in. This is simply a recognition that when you uncover a problem during testing, you have fewer options – you generally prefer to avoid spinning the board, so you try to find fixes at the box level. Of course, if the enclosure is non-conductive, this goal may not be realistic, so you will need to proceed to internal fixes.

Once you have identified potential problem areas, it is time to come up with some fixes. Here are some common remedies:

Cable Fixes
Since cables are very often part of the problem, let’s start there. Again, your options depend on your enclosure. If you have a shielded cable, the connector termination is the biggest suspect. The termination at cable shield to connector, connector to mating connector and mating connector to bulkhead all need to be well done, or the cable will leak. The termination must be circumferential at each junction – pigtail terminations and single point grounding is not acceptable. Especially note: purchased cables are rarely terminated well.

You can put in a temporary fix using copper tape to close possible gaps in the connector area. If this is not feasible, use aluminum foil to make a temporary shield over the cable shield, grounding the foil to the housing at each end. If the patch works, look for the breach in the shield – often a pigtail termination or, worse yet, the shield is grounded at only one end (and sometimes not grounded at either end).

And don’t forget to check to make sure the mating surfaces are fully conductive.

If your cable is filtered but not shielded, make sure the filter assembly is circumferentially terminated to the housing or to the cable connector, whichever is applicable.

If you don’t have a shielded enclosure, you will probably need to work on the circuit board, discussed below.

Enclosure Shield Fixes
For shielded enclosures, the principal issue is the openings and penetrations. The cable penetrations were discussed above, leaving switches, indicators, and fastener penetrations to deal with. The openings include seams, ventilation, and displays.

The penetrations are primarily a problem with ESD, and these can generally be localized by ESD testing. Discharge to switches, edges of touch panels and indicators and screw threads are prime suspects. Discharge near seams will couple energy to internal cables located near the seams. The problem with plastic enclosures is quite different. As there is no metal to which discharge can occur, the arc can only penetrate through openings in the enclosure, however small and indirect – ESD can travel surprising distances to find metal. Generally, the only option is to prevent discharge from penetrating to the internals.

For radiated emissions and immunity, seams and openings need to be closed, typically starting with the largest opening and those near cables. For immunity, you can check suspect openings with handheld radios. For emissions, sniffer probes and a spectrum analyzer will help identify possibilities.

Depending on the opening, you can close them with copper tape and aluminum foil. Closing openings like ventilators or displays may need to be covered with conductive screen or possibly perforated aluminum foil. If there are a number of suspect openings, it may be better to close up all of them by wrapping the entire box, then remove patches of foil one at a time while evaluating results.

Circuit Board Fixes
If you have no external shields to work with, your only option is to work at the circuit board, often at the board/cable boundary. Handheld radios can be used to force failures, especially at power and data cables. Sniffer probes and spectrum analyzers can help isolate an emission issue – typically by tuning to a problem frequency, then moving the probe around the board. The smaller probes are more selective but less sensitive. H-field probes work best on traces and cables, E-field probes work best on open connectors and chips. If you want to get really local, connect a high impedance scope probe to the spectrum analyzer input to get right down to the pin.

Most emission and immunity problems will be traced to the board boundary to the cables, using filters, transient protectors or a combination, whether this is radiated or conducted, emissions or susceptibility. If your problem is emissions, you have the additional option of filtering power and signals on-board, and using on-board shields.

Circuit board problems are best uncovered during pre-test, when you have some circuit board design flexibility.


Troubleshooting EMI is an uncertain process, and usually takes more than one iteration to run down even a single problem, not to mention cases where multiple problems exist.

A methodical approach can reduce the number of false trails. Start by gathering information, establishing a list of probable causes, deciding what tools and components will be needed, minimizing the system, trying fixes until results are satisfactory.  favicon




Daryl Gerke, PE and Bill Kimmel, PE
are the founding partners of Kimmel Gerke Associates, Ltd. The firm specializes in EMC consulting and training, and has offices in Minnesota and Arizona. The firm was founded in 1978 and has been in full time EMC practice since 1987.

Daryl and Bill have solved or prevented hundreds of EMC problems in a wide range of industries – computers, medical, military, avionics, industrial controls, vehicular electronics and more. They have also trained over 10,000 designers through their public and in-house EMC seminars.

Daryl and Bill are both degreed Electrical Engineers, registered Professional Engineers, and iNARTE Certified EMC Engineers.

Between them, they share over 80 years of industry experience. For more information and resources, visit their web site at


Related Articles

Digital Sponsors

Become a Sponsor

Discover new products, review technical whitepapers, read the latest compliance news, trending engineering news, and weekly recall alerts.

Get our email updates

What's New

- From Our Sponsors -

Sign up for the In Compliance Email Newsletter

Discover new products, review technical whitepapers, read the latest compliance news, trending engineering news, and weekly recall alerts.